Read in crime data and convert x/y to appropriate projection

Now create a df for the period after the pandemic

Make point maps

The crimes of interest for these maps are:

All violent crimes since 2012

All property crimes since 2012

Post-covid violent crimes

Post-COVID property crimes

Post-COVID auto thefts

Homicide heat maps

Pre-COVID homicides

Post-COVID homicides

Comparison of year-over-year census tract trends

The above plot has a clear outlier that will interfere with the fitted line. Let's try again with it removed.

There's clearly a bit of an association, but not as strong as it appeared before.

Census tract-level maps

Read in tract file

Join tract with crime data (which will help filter to only MKE tracts)

Calculate changes in crime from pre-post pandemic

Merge shapefile with data

Merge with population totals and income figures

Create per-100k rates for crimes for pre- and post-COVID

Choropleth maps

Violent crime rate

Make some histograms to get a sense of the distribution of crime across tracts, and also to get a sense of what appropriate bin ranges could be for the various colormaps.

Non-auto theft property crime rate

Auto-thefts

Auto theft hexbin plots

Compare crime rates against incomes

There is a very strong and unsuprising association between violent crime and income (above). But did the areas that saw shifts in post-COVID violent crime tend to have higher or lower income levels (below)? Not much of a relationship here.